A solid state electrolytic capacitor is made from a porous pellet of sintered tantalum powder, a dielectric tantalum oxide layer formed on the surface of the sintered tantalum powder, a solid-state conductor impregnated into the volume of the pellet, and external connections such as silver paint, etc. The tantalum forms the positive electrode of the capacitor and the solid-state conductor forms the negative electrode (also called the cathode or counter-electrode).
Manganese dioxide has been employed as the solid-state conductor in tantalum capacitors for the last 50 years. To impregnate the pellet with manganese dioxide, it is dipped into a solution of manganese nitrate. The pellet is then heated in air or steam to temperatures in excess of 200.degree. C. to pyrolyze the manganese nitrate to manganese dioxide. This process of dipping and pyrolysis is repeated several times to fill the pellet with manganese dioxide. By carefully choosing the sequence of concentrations of manganese nitrate and the pyrolysis conditions, a capacitor with a high capacitance recovery and a high volume fill of manganese dioxide can be produced.
A key property of manganese dioxide is its self-healing ability. At defective portions of the dielectric film, the manganese dioxide becomes non-conductive. This is due to the manganese dioxide transforming to a lower manganese oxide because of joule heating at the defect site. This mechanism allows capacitors with low leakage currents to be produced. It also allows small dielectric defects that occur during manufacture and use to be isolated. However, if the dielectric defect is too large, the dielectric can crack. Manganese dioxide is a powerful oxidizing agent. When it comes in direct contact with tantalum through a crack in the oxide, the capacitor can ignite, leading to destruction of the capacitor and possible destruction of other components in the circuit. It is desirable to replace the manganese dioxide with a solid-state conductor that is non-oxidizing, therefore eliminating tantalum ignition while maintaining the self-healing mechanism.
The use of tantalum capacitors in high frequency circuits has become more important. This has led to the need for tantalum capacitors having low equivalent series resistance (ESR). The best manganese dioxide has a resistivity of 0.5 to 10 ohm-cm. It is desirable to replace the manganese dioxide with a solid-state conductor that has a lower resistivity. However, many highly conductive metals and oxides do not have a self-healing ability and thus are not suitable for solid-state tantalum capacitors.
Conductive polymers such as polypyrroles, polyanilines, and polythiophenes have resistivities 10 to 100 times less than that of manganese dioxide. Since they are much less powerful oxidizing agents than manganese dioxide, these materials do not cause the capacitor to ignite upon failure. Polypyrrole was shown to have a self-healing mechanism (Harada, NEC Technical Journal, 1996).
Chemical oxidative polymerization is an effective way to impregnate conductive polymer into the pores of the tantalum pellet. In chemical oxidative polymerization, a monomer, oxidizing agent, and a dopant are reacted inside the porous pellet to form the conductive polymer. Monomers include pyrrole, aniline, thiophene, and various derivatives of these compounds. The oxidizing agent can be either an anion or a cation. Typical anion oxidizers are persulfate, chromate, and permanganate. Typical cations are Fe(III) and Ce(IV). The best dopants are anions of strong acids such as perchlorate, toluenesulfonate, dodecylbenzenesulfonate, etc. The reaction between monomer, oxidizing agent, and dopant can take place in a solvent such as water, an alcohol, a nitrile, or an ether.
Several methods have been used to get the monomer, oxidizing agent, and dopant into the porous pellet and carry out the conversion to conductive polymer. In one method, the pellet is first dipped in a solution of the oxidizing agent and dopant, dried, and then dipped in a solution of the monomer. After the reaction is carried out, the pellet is washed and then the process is repeated until the desired amount of polymer is deposited in the pellet. In this method, it is difficult to control the morphology of the final polymer. It is also difficult to control the exact reaction stoichiometry between the monomer and the oxidizing agent. Control of this stoichiometry is critical to achieve the highest conductivity polymer (Satoh et al., Synthetic Metals, 1994). Cross contamination of the dipping solutions is a problem. Since the pellet must be dipped twice for each polymerization, the number of process steps is greatly increased. The excess reactants and the reduced form of the oxidizing agent need to be washed out of the part. This adds even more process steps and complexity to the process.
In a related method, the sequence is reversed so that the pellet is dipped in the monomer solution first and the solvent is evaporated away. The pellet is then dipped in the oxidizing agent/dopant solution and the reaction is carried out. This method suffers from all the disadvantages of the previous method. In addition, some monomer may be lost in the solvent evaporation step.
In the preferred method, all components are mixed together and the pellet is dipped in the combined solution. This method reduces the number of dips and allows more precise control over the reaction stoichiometry. However, the monomer and oxidizing agent can react in the dipping bath, causing premature polymerization and loss of reactants, adding some cost and complexity to the process. This is especially a problem with pyrrole monomer and Fe(III) oxidizing agents. To partially overcome this, the dipping bath can be kept at cryogenic temperature (Nishiyama et al., U.S. Pat. No. 5,455,736). However, use of cryogenic temperatures adds considerable equipment and operational complexity to the process.
The pyrrole/Fe(III) can be replaced with a monomer/oxidizing agent combination that is less reactive. For example, 3,4 ethylenedioxythiophene and an Fe(III) salt of an organic acid may be dissolved in alcohol or acetone (Jonas et al., U.S. Pat. No. 4,910,645). With this combination, dilute solutions (less than 5% monomer) are stable near room temperature for several hours. The polymer (poly (3,4 ethylenedioxythiophene) or PEDT) can be formed by warming the solution. At concentrations greater than about 5% monomer, the components react quickly. Cooling the dipping solution can be used to retard the reaction. However, there is a lower limit to which the solution can be cooled because of limited solubility of the components at low temperatures. Addition of a nonvolatile organic base, such as imidazole, also inhibits the reaction (Mutsaers et al., EP 0615256 A2, 1994; de Leeuw et al., 1994). However, this leaves an organic residue in the pores of the tantalum capacitor that is difficult to wash out.
Because of the trend toward higher surface area (higher charge) tantalum powders, it is desirable to use more concentrated monomer/oxidizing agent/doping solutions to more effectively cover the internal surface area of the tantalum pellet in a reasonable number of dips. The instability of solutions of monomers/oxidizing agents/dopants at higher concentrations is an impediment to achieving that objective.
It is known in the art that the choice of solvent can greatly affect the reaction rate between the monomer and the oxidizing agent. For example, Myers (J. Electron. Mater., 1986) prepared polypyrrole from pyrrole and Fe(III) chloride in several different solvents. The preparation was done by mixing the components in solution and allowing the reaction to take place without evaporation of the solvent. The highest yields of polypyrrole were obtained in solvents which did not appreciably complex with Fe(III). Solvents which showed a strong exotherm on addition of Fe(III) gave very low yields.
More recently, Lebedev et al. (Chem. Mater., 1998) studied the reaction between an Ag(I) salt and neutral forms of polythiophene derivatives. The Ag(I) oxidizes the neutral polymer to an oxidized conducting form of the polymer. The counter ion of the salt becomes the polymer dopant. This reaction takes place quickly in solvents such as xylene, toluene, heptane, and chloroform. In order to stabilize such solutions, Lebedev et al. use a mixed solvent system of toluene, heptane and pyridine. The pyridine complexes with Ag(I) and prevents the reaction in solution. When the pyridine is allowed to evaporate from the mixed solvent system, the Ag(I) oxidizes the polymer to a conducting form. The system of Lebedev et al. is not suitable for use in preparation of capacitors for several reasons: the Ag(I) reactant is expensive, the Ag(O) reaction product would be difficult to remove, and the neutral polythiophenes are only soluble in low concentrations.
A mixed solvent system is also taught by Nakama (U.S. Pat. No. 5,514,771, 1996). Ion-doped poly(alkyl-substituted pyrrole) with a concentration of greater than 5% is prepared in an organic solvent and may be used as a solid electrolyte in a capacitor. Alkyl-substituted pyrrole is dissolved in a solvent comprising THF, aliphatic alcohols, or mixtures thereof An oxidizing agent such as a ferric (III) salt polymerizes the alkyl-substituted pyrrole, and water is added to separate the product by filtration. The mixed solvent system dissolves the reactants and does not prohibit the polymerization reaction.